Vision-Language Model IP Protection via Prompt-based Learning

Code release for "Vision-Language Model IP Protection via Prompt-based Learning" (CVPR 2025)

Paper

Vision-Language Model IP Protection via Prompt-based Learning (CVPR 2025)

We propose IP-CLIP, a lightweight ntellectual property (IP) protection strategy tailored to CLIP to protect Vision-Language Models.

Prerequisites

The code is implemented with CUDA 11.4, Python 3.8.5 and Pytorch 1.8.0.

Datasets

Office-31

Office-31 dataset can be found here.

Office-Home

Office-Home dataset can be found here.

Mini-DomainNet

Mini-DomainNet dataset can be found here.

Running the code

Target-Specified IP-CLIP

python Target_Specified/train.py

Ownership Verification by IP-CLIP

python Ownership/train.py

Target-free IP-CLIP

python Target_Free/train.py

Applicability Authorization by IP-CLIP

python Authorization/train.py

Citation

If you find this code useful for your research, please cite our paper:

@article{wang2025vision,
  title={Vision-Language Model IP Protection via Prompt-based Learning},
  author={Wang, Lianyu and Wang, Meng and Fu, Huazhu and Zhang, Daoqiang},
  journal={arXiv preprint arXiv:2503.02393},
  year={2025}
}

Contact

If you have any problem about our code, feel free to contact

• lywang12@126.com
• wangmeng9218@126.com